When I asked about verbs in -o-, Rafael and Joonas both brought up a wonderfully useful tool: using Perseus to find all words matching the pattern *oo.

But what if I don't want to find the dictionary forms of words? For instance, I suspect that an actual -o- verb (if one existed) would contract in the dictionary form, because that's what they do in Greek (and what first-conjugation verbs do in Latin too). I would like to search for verbs ending in *oere instead, but since the infinitive isn't the dictionary form, Perseus's tool is no help.

Is there another tool which can search for non-dictionary forms? Searching a large enough corpus would also work, though it's not ideal (since not every form of every verb would turn up in the corpus).

1 Answer 1


There is a program called WORDS by one Whitaker, apparently no longer maintained since early 2000s (his home page has disappeared, too), but the copies are still floating around. I am using it all the time. It uses some semi-empirical rules to deduct all possible forms for a word. Personally I find it very useful, and even thinking about rewriting and maintaining it.

Distributed with the program is a file LISTALL which is a list of all word forms in the dictionary that can possibly be generated. A caveat here is that the dictionary is quite large, and contains medieval Latin words and forms. There are markings in the database, but in LISTALL the distinction is lost.

As far as the technique goes, you first grep the LISTALL¹ file, then copy interesting forms and paste then to WORDS as input. This way it is possible to understand the algorithm's "idea" how the possible form was produced. For example,

$ grep -P 'oo$' listall
. . .

May of these are not verbs. It makes sense to copy the grep result and paste it to WORDS as input². Here's the output from pasting a few words into its console from the above list.

curotrophoo                    ========   UNKNOWN


eo.o                 N      2 1 DAT S M                   Always    mostfreq
eo.o                 N      2 1 ABL S M                   Always    mostfreq
Eous, Eoi  N (2nd) M   [XXXDX]      lesser
morning star; Oriental, dweller in the east; one of the horses of the Sun;
eo.o                 ADJ    1 1 DAT S M POS               Always    mostfreq
eo.o                 ADJ    1 1 DAT S N POS               Always    mostfreq
eo.o                 ADJ    1 1 ABL S M POS               Always    mostfreq
eo.o                 ADJ    1 1 ABL S N POS               Always    mostfreq
Eous, Eoa, Eoum  ADJ   [XXXDX]      lesser
eastern; of the dawn; belonging to/of/set in the morning;


epogdo.o             ADJ    1 1 DAT S M POS               Always    mostfreq
epogdo.o             ADJ    1 1 DAT S N POS               Always    mostfreq
epogdo.o             ADJ    1 1 ABL S M POS               Always    mostfreq
epogdo.o             ADJ    1 1 ABL S N POS               Always    mostfreq
epogdous, epogdoa, epogdoum  ADJ   [DDXES]    Late  uncommon
epogdo.o             ADJ    2 6 DAT S C POS               Always    mostfreq
epogdo.o             ADJ    2 6 ABL S C POS               Always    mostfreq
epogdoos, epogdoos, -  ADJ   [DDXES]    Late  uncommon
nine-eighths; (music); whole and eighth;


euro.o               ADJ    1 1 DAT S M POS               Always    mostfreq
euro.o               ADJ    1 1 DAT S N POS               Always    mostfreq
euro.o               ADJ    1 1 ABL S M POS               Always    mostfreq
euro.o               ADJ    1 1 ABL S N POS               Always    mostfreq
eurous, euroa, euroum  ADJ   [XXXDX]      lesser

So naturally none of these are verb forms, and one was not even understood by the parser. In fact, this simple example only found the same [re]boo and inchoo as the above link. But searches in the local file can be as complex as you fancy, not limited by a Web application interface.

ADDENDUM: Col. William A. Whitaker (USAF-Retired), who worked on the ADA language while at DARPA and wrote the WORDS program in his retirement, passed away in December 2010. As far as I can tell, nobody is supporting the original version at this time.

¹ The file is referred to as LISTALL in the documentation, but the actual file name as distributed is lowercase listall. I am keeping it consistent with the documentation, except in actual usage examples.

² The program accepts whitespace and newline as word separator, so that many words at once, a sentence, or even whole paragraph can be pasted to its input console.

  • Interesting, so I would grep LISTALL for the pattern I wanted? Does this handle unusual cases such as both incoho and inchoo, or alucinari and halucinari and hallucinari?
    – Draconis
    May 16, 2018 at 17:43
  • @Draconis, I added an example to the answer. As for your second question, of these hallucinari is missing, if you run cat listall | grep -Pw '(incoho|inchoo|alucinari|halucinari|hallucinari)'. The other 4 forms are there. May 16, 2018 at 19:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.